Integrating a soil temperature module into AquaCrop improves simulation of crop growth and yield under straw mulching

Introduction Straw mulching is a core practice in conservation agriculture, improving the soil hydrothermal environment, enhancing water-use efficiency, and increasing crop yield, and is widely regarded as an environmentally friendly and sustainable mulching technique. However, most existing crop models simplify or neglect mulch-induced regulation of soil temperature and water processes, thereby limiting their capacity for system-level assessment of crop growth and water dynamics under mulched conditions. This study aims to enhance the adaptability of AquaCrop to mulched environments and to evaluate the interactive effects of straw mulching on soil water–heat processes and crop growth. Methods Based on the AquaCrop framework, a soil temperature submodule (ST) incorporating key surface-cover environmental factors was developed to extend the model’s capability to simulate soil temperature dynamics. Early-season crop development was driven by cumulative thermal time in the topsoil rather than air-temperature growing degree days, yielding a coupled model, AquaCrop–ST. A maize conservation-agriculture field experiment was conducted in Northeast China using locally prevalent mulching practices, with four treatments: straw crushing mulch, straw strip mulch, whole-straw mulch, and a no-mulch control. The coupled model was evaluated against observations. Results and discussion Straw mulching significantly moderated topsoil temperature dynamics, lowering soil temperature by 1.77–4.46 °C during warm periods and reducing early-season thermal accumulation by 10–24%. The improved model reproduced topsoil temperature dynamics under mulching well, with R 2 values of 0.92–0.95, RMSE values of 0.93–1.06 °C, and NSE values of 0.88–0.94, and also captured the seasonal patterns of deeper soil temperature. Compared with the original model, canopy development simulations were substantially improved, with RMSE decreasing from 33.49–33.73% to 3.13–8.13%. The improved model also successfully captured mulch-induced phenological delay and improved the simulation of soil water dynamics and evapotranspiration.